Determinants of Facebook usage among Polytechnic Students in South East, Nigeria
Determinants of wages: Evidence from Nigeria
Transcript of Determinants of wages: Evidence from Nigeria
Determinants of wages: Evidence from Nigeria
Abimbola Olanrewaju Oni
Submitted to the
Institute of Graduate Studies and Research
in partial fulfillment of the requirements for the Degree of
Master of Science
in
Economics
Eastern Mediterranean University
August, 2014
Gazimağusa, North Cyprus
Approval of the Institute of Graduate Studies and Research
Prof. Dr. Elvan Yılmaz
Director
I certify that this thesis satisfies the requirements as a thesis for the degree of Master
of Science in Economics.
Prof. Dr. Mehmet Balcilar
Chair, Department of Economics
We certify that we have read this thesis and that in our opinion it is fully adequate in
scope and quality as a thesis for the degree of Master of Science in Economics
Assoc. Prof. Dr. Gülcay Tuna Payaslıoğlu
Supervisor
Examining Committee
1. Prof. Dr. Salih Katircioglu
2. Assoc. Prof. Dr. Fatma Guven Lisaniler
3. Assoc. Prof. Dr. Gülcay Tuna Payaslıoğlu
iii
ABSTRACT
The structure of the labor market has a significant consequence on employment
status which indirectly also serves as an important determinant of household income
and welfare. Hence this research work explored the determinants of wages and
degree of wage inequality with respect to gender, education, experience, religion, and
public versus private and self- employed versus private sectors using a recent survey
data of Nigeria General Household Survey (GHS) provided by the National Bureau
of statistics (NBS) conducted in 2010-2011 on 5,000 households from all states.
Using the Mincer type wage equation on the return to education and experience, our
findings show that men earn less than women by 16.1 percent at all levels of
education which gave similar results in the extended specifications. The coefficient
for experience variable is very small which is 0.35 percent and not significant at
conventional levels. This finding means that on-the-job training plays no significant
role in wage determination. Similarly, self-employed earn 16.4 percent less than
other employment categories in the rural area while the dummy for religion is highly
significant showing that Christians earn 23 percent higher than the base category in
the model. While the wage differential is constant at all levels of education for the
whole country, it waries with level of education for the rural area, gender gap
increasing at higher levels of education in favor of females.
Keywords: SAP, wage gap, employment category.
iv
ÖZ
Emek piyasasının yapısı, çalışanın ekonomik durumu ve dolayısı ile hanehalkı gelir
ve yaşam standardını belirlemede önemli rol oynamaktadır. Bu çalışmada, Nijerya
Ulusal İstatistik Bürosu (NBS) tarafından 2010-2011 döneminde 5,000 hanehalkı
için yapılan Nijerya Genel Hanehalkı Anketi (GHS) kullanılarak ücreti belirleyen
faktörler ve ayrıca cinsiyet, eğitim, iş tecrübesi, din, kamu sektörü, özel sektör ve iş
sahibi emek piyasası kategorilerine göre ücret ayırımcılığı araştırılmıştır.Mincer tipi
ücret fonksiyonu tahmin sonuçları, ülke genelinde, erkeklerin kadınlara kıyasla
percent 16.1 daha düşük ücret aldıkları saptanmıştır. Diğer yandan, iş tecrübesi
değişkeni oldukça düşük, 0.35 percent, olmakla beraber istatistiksel bakımdan
anlamlı bulunmamıştır. Hemülke genelinde, hem de kırsal bölgelerde kendi işinde
çalıanların daha düşük ücret aldıkları, bu oranın kırsal bölgeler için percent16.4
olduğu saptanmıştır. Ayrıca, yine kırsal bölgelerde, Hıritiyanların percent23 daha
fazla ücret aldığı bulunmuştur.Tüm ülke genelinde, kadın erkek arasındaki ücret
farkı, eğitim seviyesi ile değişmemekte ancak, kırsal bölgelerde eğitim seviyesi
yükseldikçe, farkin erkekler aleyhine arttığı tespit edilmiştir.
Anahtar kelimeler: SAP, ücret farkı, iş, emek piyasası.
v
DEDICATION
To
The Everlasting God
Through
Prof. Emmanuel Olayiwola Oni, entire family and friends.
vi
ACKNOWLEDGMENT
In a special way would like to thank my supervisor Assoc. Prof. Dr. Gulcay Tuna
Payaslioglu for her wonderful support, advice and guidance in the preparation of this
research work. Her invaluable supervision exposed me to the frontiers of
econometric research.
I am grateful to all my instructors at the department who contributed a lot to the
knowledge and skills I have gained at EMU. Besides, a number of friends and
colleagues had always been around to support me both morally and academically; I
would like to thank them as well. Ikechukwu Darlington NWAKA have stood with
me and shown what it means to have a brother. I say thank you.
To my main sponsor, Prof. Emmanuel Olayiwola Oni and other wonderful and
caring family; I remain very grateful for the moral and financial support all through
the programme. Equally, I owe quite a lot to my parents for the love and words of
wisdom thus far.
Above all to God almighty for His abundant graces, blessings and love upon me all
these while. May His name be glorified from ages to ages. Thank you Lord Jesus!
vii
TABLE OF CONTENTS
ABSTRACT ...................................................................................................................... iii
ÖZ ..................................................................................................................................... iv
DEDICATION ................................................................................................................... v
ACKNOWLEDGMENT ................................................................................................... vi
TABLE OF CONTENTS ................................................................................................. vii
LIST OF TABLES ............................................................................................................ ix
LIST OF FIGURES ........................................................................................................... x
1 INTRODUCTION .......................................................................................................... 1
1.1 Aim of the study....................................................................................................... 3
1.2 Structure of the study ............................................................................................... 4
2 THEORETICAL FRAMEWORK .................................................................................. 5
2.1 Human Capital Framework ...................................................................................... 5
2.1.1 Schooling as Human Capital Investment .......................................................... 6
2.1.2 On-the- job Training as Experience .................................................................. 7
2.1.3 Ability ............................................................................................................... 8
2.2 Empirical Literature Review .................................................................................... 9
3 NIGERIAN LABOUR MARKET ................................................................................ 12
3.1 Structure of the Nigerian Labor Market ................................................................. 12
3.2 Nigerian Wage Review Policies ............................................................................ 15
3.2.1 Wage Inequality in Nigeria ............................................................................. 16
3.3 Nigerian Educational System ................................................................................. 16
3.4 Empirical Literature on Nigeria ............................................................................. 18
viii
4 DATA AND METHODOLOGY .................................................................................. 22
4.1 Data and Descriptive Statistics .............................................................................. 22
4.1.1 Chi-square Goodness-of-Fit Test .................................................................... 29
4.2 Methodology .......................................................................................................... 31
5 EMPIRICAL FINDINGS ............................................................................................. 32
5.1 The Basic Model: Return to Education and Experience ........................................ 32
5.2 Extended Model ..................................................................................................... 35
6 SUMMARY AND CONCLUSION ............................................................................. 42
6.1 Conclusions ............................................................................................................ 42
6.2 Recommendations for Further Studies................................................................... 43
REFERENCES ................................................................................................................ 45
ix
LIST OF TABLES
Table 1: Basic Labor Market Indicators in Nigeria………........................................11
Table 2: Detailed Summary Statistics of logarithmic wages......................................21
Table 3: Distribution of Education by Gender............................................................24
Table 4: Years of Education.......................................................................................25
Table 5: Wage Earners by Region and Gender...........................................................25
Table 6: Wage Earners by Category of Employment and Gender..............................26
Table 7: Return to Education and Experience............................................................29
Table 8: Return to Education by Gender....................................................................30
Table 9: Return to Education by Gender at Average Education Level.......................31
Table 10: Extended Model compared with Basic Model............................................33
Table 11: Determinants of Wages for the Rural Sector..............................................35
x
LIST OF FIGURES
Figure 1: Histogram of wages and logarithmic wages………….………..………….21
Figure 2: Box Plot for logarithmic wages...................................................................22
Figure 3: Logarithmic Wages for Males against Females..........................................23
Figure 4: Logarithmic Wages with respect to Educational Level Attained................24
1
Chapter 1
1 INTRODUCTION
Since independence in 1960, the Nigerian economy despite its vast number of natural
resources such as crude oil has experienced some economic instabilities leading to
unemployment and inefficient allocation of resources. Over the years, agriculture
which has been the main stay of the economy has been neglected recording no
visible development. In 1960’s, agriculture contributed about 65-70 percent to the
federal government revenue and this was significantly more than what other sectors
contributed at the time. In subsequent years, low activity of the agricultural sector
has led to decrease in its contribution to the GDP. This can be said to be as a result of
discovery of crude oil which now contributes larger share to the country’s economy.
Nigeria being the largest economy in Africa with estimated population of about 162
million is ranked the world’s seventh most populated country in the world.
According to the 2011 Central Bank of Nigeria (CBN) Bulletin, about 38 percent of
females lack formal education, increasing the proportion of illiterates and
inefficiency in economic activity. An estimate of about 22 percent of the male
counterpart lacks formal education. The unemployment rate in Nigeria has been on
the increase since the neglect of agricultural sector: the National Bureau of Statistics
of Nigeria reported an increase to 23.9 percent in 2011 from 21.1 percent in 2010.
Aminu (2011) reported that the Nigerian unemployment rate averaged 14.6 percent
2
between 2006 and 2011 which reached the highest value of 23.9 percent in
December of 2011 and a recorded the lowest of 5.3 percent in December of 2006.
In 1986, the Structural Adjustment Program (SAP) was introduced to restructure and
diversify the productive base of the economy in order to reduce dependency on the
oil sector, promote the non-inflationary economic growth and finally to achieve
fiscal and balance of payments viability Analogbei (2000). Nigeria has opened up
trade and has undergone capital market liberalization in parallel to the globalization
trend in the world. Like other developing countries, this was expected to affect the
Nigerian economy positively. For example as Tokarick (2008) reported, agricultural
trade liberalization has positive welfare effects through removal of tariffs and
subsidies. Accordingly, while the removal of tariffs leads to increased demand for
the product, thus increasing prices, the removal of subsidies would shift resources
from the agricultural sector to other sectors that also would affect prices upward
leading to welfare gains for the net exporters of agricultural products. However, the
Nigerian economy still suffers from wage inequality between the agricultural sector
and oil sector which received more attention after adoption of the SAP program. The
reasons can be explained by three main factors in summary; first, the positive
response of the agricultural sector to the SAP program could not be sustained due to
mainly unsatisfactory agricultural financing and small scale of farmers, infrastructure
and management problems, also because of the fact that the manufacturing sector
was still highly dependent on import of foreign inputs, and the oil sector was mainly
owned by foreigners.
3
After 1986, new policies were implemented by the federal government to improve
the labor market as well. This trend was established to improve the wage rate of the
public sector and enhance the effectiveness of people Aminu (2011). However, this
led to migration into the cities, as government wages are favorable compared with
income opportunities in the rural areas. This has led to wage gap between rural and
urban, public and private sector and protected and unprotected sectors.
The structure of the labor market has a significant consequence on employment
status and it serves as an important determinant of household income and welfare.
Like most labor markets in developing countries, the Nigerian labor market is not
growing as fast as the population and this has increased poverty over the years. In
this respect, it is important to explore the features of the Nigerian labor market to
explain wage inequalities with regard to individual characteristics of labour, public
and private sector and self-employers. Wage differential with respect to gender is
another issue which may be contributing to wage inequality in Nigeria for which a
Gender Equality Legislation was enacted by the 2007 Gender Equality Duty
Aromolaran (2006), in order to address the issue.
1.1 Aim of the study
So long as agriculture is the highest employer of labor, second largest GDP
contributor and the major means of Nigerian people livelihood, the aim of the
research is to introduce a model that will investigate the determinants of wages in
Nigeria covering 36 states both rural and urban during a recent time in 2011. The
wage equation is of Mincerian type, however, in addition will include other
individual and regional characteristics such as gender, religion, employment
categories and state of residence. In this respect, the paper will explore the degree of
4
wage inequality with respect to gender, religion, and public versus private and self-
employed versus salary workers using the recent survey data of Nigeria General
Household Survey (GHS) provided by the National Bureau of statistics (NBS)
conducted in 2010-2011 on 5,000 households from all states.
1.2 Structure of the study
First, theoretical literature of wage function is presented in chapter two. Description
of the variables used within the human capital theory will be explained in this
chapter followed by the empirical literature review showing other researchers
contribution. The Nigerian labor market will be analyzed in the third chapter as we
proceed. Information on the survey data used and descriptive statistics on key
variables and the estimated model will be presented in chapter four. Chapter five
will present the estimation results while chapter six will be on conclusion with
remarks.
5
Chapter 2
2 THEORETICAL FRAMEWORK
This chapter will present the theoretical framework for wage determinants which is
based on an extended version of Mincerian human capital model. Furthermore,
major work contributing to the literature with different empirical findings will also
be summarized.
2.1 Human Capital Framework
The literature is well established for determining factors contributing to wages
through human capital theory. Human capital formation is understood to be
productive capabilities of human generating income. Therefore, the determinants of
earnings have played a major role in human capital theory. Thus, this formation is
related with investment in human and capital development, Thurow (1975). For
instance, in (1776), Adam Smith emphasized the importance of education at different
points of life. Later, an important pioneering work by Jacob Mincer (1957, 1958,
1962) contributed to the formation of a statistical wage equation which focused on
individual characteristics affecting earnings which is simply expressed as
= + eq. (1)
where is the natural logarithm of wages or earnings and is the vector of
explanatory variables on individuals’ quantitative skills which would basically
include education and experience, and is the disturbance term.
6
Differences on wage structure and factors influencing wage function have generated
different opinions and viewpoints from different researchers over the years. These
studies have considered also other factors related with wages such as environment,
family background, regional or policy factors in addition to education and
experience. (Mincer and Polacheck, 1974 and Mincer, 1974). In this respect, the
modern capital theory extends the traditional wage equation to include other personal
characteristics such as gender, race, religion, employment and or geographical
characteristics providing richer specifications to study discrimination and wage
differentials. The theoretical background and their interpretations as contribution to
basic determinants of wages will be summarized below.
2.1.1 Schooling as Human Capital Investment
Human capital which is accumulated knowledge and skills of an individual increases
his or her productivity. In the literature, education is accepted as the main factor
improving productivity of a worker. However, additional schooling involves
opportunity costs of foregone earnings plus explicit costs such as tuition fees. In this
respect, return to education gains a substantial role for the decision of an individual
about the level of education that should be attained. Three propositions are discussed
in the literature as supply-side, demand-side and market equilibrium. Considering
the labor supply side, in order to be attracted to further schooling and forgo present
earnings, an individual must be compensated with increases in life time earnings.
Second, on the labor demand side, the marginal productivity of an individual with
more schooling must be considered to be more than that of a less schooling
individual. Finally in the long-run, market equilibrium of each schooling category
must be maintained such that the workers will not alter their education level Berndt,
(1991).
7
Another view developed by Arrow (1973) and Spence (1973, 1974) is the screening
hypothesis of education: firms consider diplomas as signals about the capabilities and
knowledge of workers rather than education being directly related with their
productivity.
2.1.2 On-the- job Training as Experience
Another way of accumulating skills other than formal education is attained by on -the
-job training which may include both formal training programs by firms or informal
“learning by doing”. (Becker; 1962, 1964 and Mincer;1962, 1974). Therefore,
according to the human capital theory, human capital can be increased by additional
job experience. In other words, years of labor force experience is a measure of a
stock of human capital attained by on-the- job training. However, years of labor force
experience is not directly observed from survey data and instead, computed as age
minus years of schooling minus six years assuming that all individuals start primary
school at six years of age. One other important issue regarding experience is that
earnings or wages should not be constant after leaving school but peak at about
midlife and then start falling representing diminishing returns. Within this setting, the
wage equation is formulated by addition of the experience variable in quadratic form
as
= + + + 2 + eq. (2)
where is expected to be positive while to be negative. Similarly, returns to both
formal education and work experience decline with age. This is because, as argued in
the literature the optimal time for investing in education is at early ages as the
opportunity cost of foregone income will be lower and likewise, tendency to add to
human capital by experience will also fall by age. Also, human capital depreciates as
time passes which implies that there is tendency to forget at older ages. Following
8
this argument some studies use age as an instrument to experience especially for
small samples as in Card (1993).
2.1.3 Ability
Another notable aspect of human capital theory involves the differences in abilities
of different individuals. Ability is defined as the capacity of accumulating skills prior
to education or training which is invariant of time. In this case, the benefit attributed
to schooling by an individual who is brilliant will be more than for an individual who
spend more time in acquiring the same level of knowledge. In other words, the more
brilliant individual is more likely to invest in education as it will have better
opportunities such as learning quickly and thus, accumulating human capital rapidly
or obtaining scholarships. Also, more educated individual will be trained at less cost.
Thus, one implication of schooling in human capital theory is that there exists a
correlation between ability and the years of schooling (Becker 1962). This implies
that more educated individual should gain more from on-the-job training. This fact
can be incorporated into the wage equation by adding an interaction term of the
education variable with the experience variable to obtain
= + + + 2 + eq. (3)
in which Thus marginal return to experience will be given by the first
derivative of log wages with respect to experience,
and ability which cannot be
observed directly. In this case, if ability is missing in a wage regression equation, it
will lead to omitted variable bias due to the correlation between ability and education
where education variable enters into the model to be estimated. One proposal to
solve for the problem of omitted variable bias is by Griliches and Mason (1972) who
uses IQ tests as a proxy variable for ability. Their findings showed no correlation
9
between ability and schooling indicating that omitted variable bias was not a
problem. Another approach to capture the effects of ability is to introduce other
family background information as control variables. (Taubman, 1976a, 1976b; Lam
and Schoeni,1993). This issue can also be addressed by using panel with fixed effects
in the case when one can assume that individual and family background
characteristics do not vary over time.
2.2 Empirical Literature Review
Empirical literature on determinants of earnings is vast. It is worth noting here that
various researches found varying results. First of all, empirical findings had shown
that the returns to education on wages can vary substantially. For instance, Thurow
(1975) using 1972 data found that 28 percent of people with university education had
negative returns to education while only 24 percent had positive return. On the other
hand, Hanoch (1967), found a decreasing marginal rate of return to education.
However, in general, researches have also shown that education reduces
unemployment and increases competition resulting in a good wage rate Griliches and
Mason (1972). The search for better education and higher wages has caused the
neglect of the agricultural sector especially in developing countries Steier (1987).
Many people migrate from the rural areas to the urban areas for better educational
facility and seek employment in more developed sectors.
Studies of Lindauer et. al. (1988), Bennel (1981) investigated the wage difference in
developing countries such as Kenya, Ghana and Nigeria. Their evidence showed the
widespread decline of lifetime real earnings in the manufacturing sector of Africa
over the last decades casting doubts on the importance of institutional factors causing
wage rigidity in this sector.
10
Not so many researches have been conducted on developing countries; in fact, the
results of estimates from different scholars have not been uniform. This may be due
to the unstable pay structure that characterizes the public and private sector income
in these countries. Some studies such as Lindauer and Sabot (1983), House (1984),
Terell (1993), Filmer and Lindauer (2002); Boudarbat (2004), Hyder ans Reilly
(2005) have estimated different levels of income for public sector workers in
developing countries. Other studies by Corbo and Stelcner (1983), Van Der Gaag
(1989), Bedi (1998) found out that the private sector workers enjoyed higher level of
income compared to the public sector. This un-reconcilable difference was the bases
of the study by Mohan (1986) in Columbia and Samarrai and Reilly (2005) in
Tanzania. According to the finding of these researches it was concluded that there is
no statistically significant difference between wage premium of private workers and
public workers.
Besides this three classes of studies, another important study that contributed
enormously to wage earning determinant was by Steier (1987), conducted in
Venezuela who concluded that the average income of public sector workers were
higher at all levels of education except post-secondary in 1975 and in addition, the
public wage was lower at all educational levels in 1984. In 1975, income
contribution by additional schooling and experience were said to be higher in private
sectors.
Less work investigated the wage determinants in developing countries. One
important study by Psacharopoulos (1985) which studies return to education for 60
countries which included both advanced and developing countries reported that
11
return to education is highest for primary education, the education of women and
countries with lowest income per capita confirming earlier findings. He also reported
that especially in developing countries, the average rate of return to education for
women exceeded that for men which were estimated using Mincerian type wage
equation. He explains this being due to the fact that more educated women
participate in the labor force. This study will investigate the determinants of wages in
Nigeria.
12
Chapter 3
3 NIGERIAN LABOUR MARKET
3.1 Structure of the Nigerian Labor Market
Based on the labor market statistics, in 2010, about 57.43 percent of labor force in
Nigeria comprises of the males while the other 42.6 percent represents females. In
2011, these figures are 42.48 percent for female category and 57.52 percent for males
respectively. The labor force participation rate as percent of female population
between ages 15 and 64 is 48.1 percent for 2010 is almost the same in 2011 which
was recorded as 48.2 percent. These statistics are presented in Table 1 below.
Table 1: Basic Labor Market Indicators in Nigeria
Men Women Total
Labor Force (2010) 28,546,479 21,160,084 49,706,564
Percentage Labour
Force (2010)
57.43 42.57
Labor Force (2011) 29,446,016 21,746,640 51,192,657
Percentage Labour
Force (2011)
Labor force
participation rate
percent (ages 15-64)
(2010)
Labor force
participation rate
percent (ages 15-64)
(2011)
57.52
51.9
51.8
42.48
48.1
48.2
55.7
55.9
Source: World Bank data base; World Development indicators.
13
Similar in many developing countries, the Nigerian labour market is observed to be
heterogeneous in nature with a given wage differences across various sectors of the
economy, Ogwumike, et. al. (2006). Based on World Bank data base, world
development indicators, in 2010, about 49 percent and 51 percent are living in the
urban and rural areas respectively. Similarly, for 2011 this data is 49.6 percent and
50.4 percent respectively.
Regarding public sectors, out of about 1.1 million jobs created in 2010, of which
about 9.4 percent was seen in the public sector while 53.6 percent in the informal
sector and 0.37 percent due to the formal sector, (World Bank 2010). This confirms
ILO estimates on the burgeoning and increasing number of the informal settlements
in the developing countries. According to the National Manpower Board (1998), the
Nigerian labour market is classified into seven different categories which include: the
employer, self-employed (in the agricultural sector), self-employed as a trader, self-
employed in other sectors, employed wage earners in the private firms, employed
wage earners in the public sector and paid apprentice see Ogwumike, et. al. (2006).
In addition, Federal Office of statistics (2001) included the informal sectoral
employment comprising of the unpaid family worker and own business enterprises.
This is further classified as one of the greatest sources of risk and one of the channels
into the poverty trap Ogwumike et. al. (2002). This is basically seen from those paid
employees or low skilled individuals who are easily susceptible and disadvantaged in
the event of a shock.
At the same time Nigerian labor market has been characterized as being regional and
highly immobile. Okigbo (1991) argued that, Nigeria’s labor market is similar to
14
other developing countries which consist of formal and informal sectors. These
dualistic characteristics are the basis of the segmentation theory, Leontarridi (1998).
Here, private and public sectors makes up the formal sector while the informal sector
consists of urban informal, rural oriented including unpaid family workers.
According to Becker, (1993), informal labour market is unregulated and often called
the underground economy which absorbs retrenched workers due to downsizing and
economic adjustment policies.
In 1986, the introduction of Structural Adjustment Program (SAP) in Nigerian labor
market led to several structural changes. Noticeable among these introduced policies
are the large scale reduction of public sector workers and the encouragement of the
growth of private sector which was an alternative employment choice for the
unemployed, Nwabugo (2011). The private labor market is a collection of privately
owned mostly multinational companies and national entrepreneurs. Macroeconomic
policies have been introduced to contain these changes especially the problem of
unemployment. In line with this policy framework, some of the programs introduced
include National Poverty Eradication Program (NAPEP), Loan Scheme and National
Directorate of Employment (NDE), Small and Medium Enterprises (SMS), (CBN
Bulletin 2011).
The public sector downsizing program has been fundamentally the latest
development in the Nigerian labor market. In 2001, the Nigerian government reduced
the amount of government workforce to ensure efficiency among public sector
workers. This reduction at federal level in public sector workforce has not been easy
to implement due to its fiscal implication problems, that is, such a program can
15
wreck the country. In order to improve the policy, a number of strong policies were
made to confirm the need to reduce the federal workforce. The implementation and
execution of this policy has been successful according to Aminu (2011).
3.2 Nigerian Wage Review Policies
Wage policy in Nigeria has been reviewed by different commissions dating back to
the first Hunts commission in 1934 and last in Justice Alfa Belgore committee in
2009/2010. The minimum wage rate in 1981 was fixed at 125 Naira per month by an
Act or Parliament but at the introduction of the SAP in 1986, the government issued
the National Minimum Wage (amendment) order, which bridged the 1981 Minimum
Wage Act. In 1987, this amendment was rescinded owing to labour protest against it
in major cities across the country. The aftermath of the implementation of the
recommendation of the committees led to an increase in general price level, causing
a jump in the nominal wage rate. This also led to a reduction in workers’ purchasing
power. Minimum wages are also set in Nigeria through Government Decrees (this
was used during military regimes), Acts and Legislation. For example, in 1992,
President Ibrahim Babangida approved a 45 percent increase in wages for Federal
Civil Servants.
In September 1998, the Federal Government of Nigeria issued a directive to increase
the nation’s minimum monthly wage from 363 Naira to 3,000 Naira. This also raised
earning of all other categories of employees. Furthermore, in 2000, the Federal
Government again raised the minimum wage rate from 3,000 Naira to 5,500 Naira
per month. Also it aided a lot in sustaining the employees of the civil service
compared to that obtainable in 1991. Again, President Goodluck Jonathan in August
2010 set the minimum wage at 18,000 Naira which was implemented from January
16
2011. Nigerian workers have fought and struggled for improved wages and national
minimum wage legislation Aromolaran (2006).
3.2.1 Wage Inequality in Nigeria
The United Nations Development Program (UNDP (2009) reveal that inequality in
Nigeria remained high, one of the highest in the world (0.49) which is 65 percent of
assets in the hands of 20 percent of population. Nigeria was rated as the most wage
unequal country in the world despite all the vast resources. Even the high level of
poverty in the country is partly a feature of the high rate of wage inequality and
differential access of the basic infrastructure, jobs and training opportunities, and
education. The UNDP (2009) indicates that Nigeria’s gender earning is one of the
highest in the world and that the females are in the higher paid jobs and more
powerful positions. Yet, it seems that Nigeria still has a lot to do in order to correct
the problems of wage inequality.
3.3 Nigerian Educational System
The educational system in Nigeria is viewed as a tool to achieve economic
development and thus reflects government policies so far in repositioning the sector
Amaghionyeodiwe (2006). As from 1985, the primary education was free called the
6-3-3-4 system reflecting primary education of 6 years, post primary education for 3
years of junior secondary school, 3 years of senior secondary school and 4 years of
tertiary education level. In 1999, a Universal Basic Education (UBE) scheme was
launched by the Obasanjo’s administration which targeted to increase adult literacy
rate from 57percent in 2003 to 70percent by 2010, (FRN, 2000) and provide
education by all by 2015, Edho (2009). The program was to make the primary
education free and compulsory for 9 years of primary and junior secondary education
17
which was also referred to as 9-3-4 system of education. The system was monitored
by Universal Basic Education Commission (UBEC) therefore, the UBEC law section
15 defines UBE as early childhood care and education.
The secondary years of schooling is spent in the duration of six years, which are 3
years of JSS (Junior Secondary School) and 6 years of SSS (Senior Secondary
School). The Federal Republic of Nigeria consists of 36 states and Federal capital
territory. There are about two Federal Government secondary schools in each of the
states. These colleges are funded and managed by the Ministry of Education by the
Federal Government. The other secondary schools are owned and governed by state
government private institutes. State-owned secondary schools are funded and
controlled by each state government and are not the same as the Federal government
colleges. Although education is supposed to be “free” in the majority of the state
owned institutions, here, students are required to purchase books, uniforms and pay
for miscellaneous things costing them an average of thirty thousand naira ($200) in
an academic year (US Embassy, 2012).
One of the main problems that are faced is the “closeness to school”. Most students
that cannot afford to go to a private school and leaving far from the Federal
Government Colleges and State-owned secondary schools eventually doesn’t have
the opportunity to attend school. Most of the tertiary education is government owned.
Nigeria has a total number of 128 universities registered by NUC (Nigerian
University Commission) among which state government and federal own 38 and 40
respectively while the rest are owned privately. The minimum university age of
entering the university is 18 years.
18
3.4 Empirical Literature on Nigeria
Using data from the GHS of 1996 -1999, Aromolaran (2006) analyzed the returns to
schooling using the Mincer wage equation. Comparing wage earners across both
private returns and self-employed, findings show that, the average years of schooling
for women in the private sector was quite higher than that of the males. However,
while using the self-employment regression model, it was observed that male had
higher levels of schooling. This results is consistent with that of Schultz (2001), and
Psacharopolous and Patrinos (2002).
Ogwumike et. al. (2006) studies the determinants and distribution of earnings
including income inequality in Nigeria using GHS for Nigeria. Their study applied
several econometric and descriptive methodologies such as Gini coefficient, Theil’s
Entropy Index, ordinary least squares technique, Heckman’s two-stage selectivity
bias correction procedure and Tobit model for estimating labor force participation.
Their findings show that inequality exists in paid employment than in self-employed
in Nigerian labour force. Also, paid employed women earn higher than men in the
same employment segment, while men in self-employed sector earn higher than their
female counterpart. They further concluded that earnings in the rural areas are greater
than that of the urban agricultural workers.
Ikhizama and Lawal (2008) examined the role of woman scientists in agricultural
development in the southwestern part of Nigerian using job related attributes of
women. A sample was drawn randomly on 73 woman scientists while the sampled
data was analyzed using frequency distribution and percentages. Their findings show
that women are highly integrated into various agricultural fields. However, they
19
recommend that such findings can be improved on if more women are actively
involved in various research fields.
Temesgen (2008) investigated the major determinants of establishment levels and
gender pay gap with a particular focus on analyzing the effects of labor market
institutions. He used two-stage regression model on analyzing the varying impacts
of labor market institutions and firm level characteristics on gender wage gaps in the
Nigerian urban labor market using information from worker and establishment levels
through an administered survey data. Findings show that labor market institutions
such as unions, and firm characteristics such as ownership, affect the level of gender
wage inequality at the firm level. It was also found that unions have significant
influence on firm level gender wage gaps in Nigeria, where they are historically
known to be strong.
Aminu’s (2010) research was based on determinants of earnings in wage
employment in Nigeria. Using probit model he found that the level of education
attained by employed household members either male or female is the main
determinant of labour market participation. He, therefore, concludes that the main
determinant of wages includes the educational levels of the labour force,
geographical location and experience of the individuals in the household. Return to
education and experience is not the same for male and female and the vary across his
model findings.
Fabusoro et. al. (2010) analyzed the “Forms and Determinates of Rural Livelihoods
Diversification in Ogun State, Nigeria” among 320 rural households. Their data were
20
all placed on hierarchical regression and a Simpson Index for diversification.
Findings show that nonfarm activities contribute about 63percent to income of the
households, while, the Simpson Index indicates a moderate diversification where
farm and nonfarm activities complements each other. Also education, size of
households and income were significant in predicting diversification.
Aminu (2011) while using Mincerian human capital model, investigated the effects
of the wage review process of 1998 on private and public pay differentials. His
empirical finding on urban male employees shows that, public sector was
disadvantaged in pay-gap at 6.78percent, but however enjoyed a higher wage of
35.07percent after the wage review process. Overall, public sector workers had
higher earnings after the wage review process than their private counterpart.
Ugochukwu and Chijioke (2011), examined the prospects and challenges of
productive employment in Nigeria within the macroeconomic policy framework. By
using a recursive structural Vector Autoregressive model, they found that any
increases in monetary policy rate to cut down on inflation have a depressing impact
on the economy. The study suggested that a more flexible inflation rate increased
money supply, access to credit and a modest but upward adjustment to capital and
recurrent expenditure have greater potential in accelerating GDP growth and for the
attainment of full employment and poverty reduction in Nigeria.
Jonah and Yousuo (2013), in their study analyzed the impacts of wage differentials
on labour turnover by using a logit model. Their analysis comprised 840 employees
of both private and public workers. Their findings supported the standard inverse
21
relationship between wage differential and the labour turnover. Hence, as state
workers’ wages increased, the probability of a State worker leaving to Federal civil
service indicated to fall by 0.29. They therefore recommended for a unified salary
structure with no disparity in the country.
Matthew and Azuh (2013) examined whether or not the trade liberalization process
has any effect on both the reduction in the wage differential between registered and
non-registered workers while using secondary and primary data in Nigeria. Findings
confirmed that the fall in the wage gap between registered and non-registered
workers in the manufacturing sector was affected by trade-related variables,
particularly, by the import penetration ratio. However, no robust evidence was
provided that trade liberalization had a substantial effect on the decrease in the
proportion of registered workers.
Also in a co-authored paper, Ogundari and Aromolaran (2014) used the Double
Hurdle (DH) model and Quantile regression (QR) to estimate the effects of education
on household welfare in Nigeria. Their findings showed that return to education was
more pronounced in the tertiary level of education than in both the primary and
secondary levels. They found that the impacts of education in the labour market
earnings at primary, secondary, graduate and post-graduate levels were explained by
about 3.1 percent, 1.7 percent, 8.5 percent and 1.7 percent respectively in the urban
areas.
22
Chapter 4
4 DATA AND METHODOLOGY
4.1 Data and Descriptive Statistics
The data used was sourced from the most recent General Household Survey (GHS)
conducted in 2010/11 by the Nigerian Bureau for Statistics (NBS), Nigeria. This
comprises of 22,000 individuals including household head, spouse, relatives and
dependents of the same household. Total number of households interviewed was
5,000.
However, for the main aim of this study the sample size was restricted to individuals
between ages of 17 to 60 of the working age population. Further to this, the sample
was restricted to include only the household head and spouse reporting positive
earnings rather than all family members with positive earnings. The reason is to
avoid including part time workers, if any, which most probably are children at school
who might be employed only temporarily in support of family income. On the other
hand, the justification for excluding none-wage earners is to avoid missing
observations in the sample data set. After these adjustments, the sample consisted of
2,469 observations which were 1,364 females and 1,105 males who have been
strictly household head and/or spouse reporting positive earnings by the time of the
survey conducted.
23
Figure 1: Histogram of wages and logarithmic wages.
Figure 1 presents the histogram of wages on the left and logarithmic wages on the
right which is relatively normal, only mildly negatively skewed, as compared with
the distribution of wages which is highly positively skewed as expected. The detailed
summary statistics for logarithmic wages are provided in Table 2 below.
Table 2: Detailed Summary Statistics of logarithmic wages.
Percentiles Smallest
1%
5%
10%
25%
6.9078
7.6009
8.2940
9.012
5.2983
6.2146
6.2146
6.2146
50% 9.798127
Largest
75%
90%
10.4631
11.0349
13.4871
13.6693
99% 12.5426 13.7101
Std. Dev. 1.1535
Mean 9.7249
Variance 1.3305
Skewness -0.1483
Kurtosis
No. obs.
3.4236
2469
0
1.0
e-0
52.0
e-0
53.0
e-0
54.0
e-0
5
De
nsity
0 50000 100000 150000 200000 250000W
0.1
.2.3
.4.5
De
nsity
6 8 10 12 14lw
24
As seen from Table 2, the log wages vary substantially across individuals, minimum
being 5.29 and maximum of 13.71. The median value is 9.798 greater than the mean
of 9.72 which shows that the data is negatively skewed as confirmed by the skewness
coefficient of -0.148. Kurtosis is greater than 3. The distribution of the dependent
variable in logarithms is thick tailed and only slightly left skewed.
Figure 2 displays the box plot for the logarithmic wage data where the box represents
the interquartile range from the first and the third quartile representing the majority
of wages. The horizontal lines are the whiskers calculated as upper (lower) quartile
plus (minus) 1.5 times the interquartile range or the maximum (minimum) value of
the data if smaller (larger). Therefore, data falling outside the whiskers that are
shown by dots represent the outliers within the sample data.
Figure 2: Box Plot for logarithmic wages
As observed from the box plot in Figure 2, there are extreme values for the
logarithmic wages taking values slightly above 12.6 and below 6.9 approximately.
The outliers can be identified specifically as;
Outlier above Quartile 3: 10.46 + 1.5(10.46 – 9.01) = 12.64
Outlier below Quartile 1: 9.01 – 1.5(10.46 – 9.01) = 6.84
68
1012
14lw
25
Therefore, log wage values above 12.64 and below 6.84 are identified to be outliers
within the sample which can distort estimations since they are not representative of
the whole sample data. In this respect, a trimming process is further needed to take
out the individuals with extreme log wage values. These include individuals earning
approximately 1,000 and less than 1,000 Naira and 300,000 and above 300,000 Naira
per month.
For the whole sample data, the box plot of males versus females with regard to log
wages are presented in Figure 3 where the gender dummy variable is defined as
taking the value of 1 for males and zero otherwise.
Figure 3: Logarithmic Wages for Males against Females
The box plot in Figure 3 clearly indicates that males earn less than females, in
general, while the extreme values on the lower bound for females are more than
those for males. However, the extreme values on the upper bound are more for males
than females.
68
1012
14lw
0 1
26
The logarithmic wages of individuals sampled with respect to educational level is
displayed in Figure 4 below where education is categorized as primary education,
secondary education and university education.
Figure 4: Logarithmic Wages with respect to Educational Level Attained
Figure 4 indicates that higher the level of education is, higher the salaries are and a
large wage gap occurs especially between university graduates and non-university
graduates. Also, it is observed that only minority of the sampled individuals have
university education. In order to analyze the distribution of education by gender, the
statistics for educational levels of individuals as primary (6 years) secondary (12
years) and tertiary (14 years) and university level of education (16 years) and post
university (more than 16 years) are presented by gender in Table 3.
68
1012
14lw
0 1
0 1 0 1
0 1
27
Table 3: Distribution of Education by Gender.
Educational Level Male Female Total
Freq. % Freq. % Freq. percent
Illiterate 42 1.7 30 1.2 72 2.9
Primary (6 yrs.) 342 13.9 339 13.7 681 27.6
Secondary (12 yrs.) 388 15.7 258 10.5 646 26.2
Tertiary (14 yrs.) 196 7.94 143 5.80 339 13.7
Tertiary-Univ. (16 yrs.) 156 6.32 79 3.20 235 9.5
Post University 38 1.54 12 0.48 50 2.0
Total 1,162 29.8 861 24.31 2,023 68.5
Table 3 shows that almost 54 percent of all sampled individuals attained primary and
secondary education of whom 29.6 percent are males and 24.2 percent females.
People having tertiary and university level education are only about 24.72 percent of
whom about 14.6 percent are males and 9 percent are females. The numbers of
individuals who have no education at all or have post university education are very
few which make up about 5 percent of the sample. This means that only 63.5 percent
of the sample has attained either primary, secondary, tertiary or university education.
In this respect, Table 4 presents the detailed years of education of the sampled
individuals. Accordingly, some individuals have some years of education although
they do not earn a diploma. The mean years of schooling is computed as 9.9 years
with a standard deviation of 4.4 years for the sample of 2,469 individuals.
28
Table 4: Years of Education
Table 5 summarizes information regarding regional distribution of wage earners of
household head and or spouse by gender. Accordingly, 1,938 individuals are
employed in the rural area of whom 1,111 are males and 827 are females. These
statistics indicate that most of the sampled individuals are in the rural area most of
which are males.
Table 5: Wage Earners by Region and Gender
Gender Rural Urban Total
Male 1,111 253 1,364
Female 827 278 1,105
Total 1,938 531 2,469
Table 6 presents data summarized by gender and category of employment as public,
private and self-employed. Other categories are ignored as their proportion is
negligibly small as can be derived from the table which constitutes merely 7 percent
of the sample. Table 4 clearly shows that most of the sampled individuals are self-
employed; being 67.3 percent and the distribution by gender is almost equal.
However, the public sector which appears as the next popular category of
employment with about 25 percent of employment in total constitutes of 402 male
Education
(Years)
Females
(Average)
Males
(Average)
Total
(Average)
0 30 42 72
1 – 6 73 73 146
7- 12 65 80 145
13 – 21 35 58 93
Total 203 253 456
29
employees that is twice as much as females. The share of the private sector is small
in which the majority are males.
Table 6: Wage Earners by Category of Employment and Gender
Gender Public Private Self-employed Total
Male 402 177 802 1,381
Female 215 57 746 1,018
Total 617 234 1,548 2,399
In summary, the descriptive statistics indicate that most sampled individuals have
primary and secondary education, mostly living in the rural area and working as self-
employed. Also, the higher the level of education is higher the level of wages earned
and thus, females, in general, earn more than males. However, these results are only
based on descriptive statistics and have to be formally modelled by a wage equation
to identify the determinants of wages for the sample data. As explained earlier, the
data is trimmed further to delete the outliers on the logarithmic wage data;
individuals with extreme log wage values. These include individuals earning
approximately 1,000 and less than 1,000 Naira and 300,000 and above 300,000 Naira
per month that reduced the sample size to 2,408 number of observations of which
1,079 were females and 1,329 males.
4.1.1 Chi-square Goodness-of-Fit Test
Before proceeding further, we need to check whether the sampled data represents the
population proportions of males and females in the labor force. This issue plays an
important role at the stage of modeling the wage equation. In order to look into the
question, we applied a nonparametric goodness-of-fit test using chi-square
distribution. Our motivation for applying the chi-square test is to compare the
30
distribution of the original labour force by gender at the national level with the
distribution in our sample. The ratios of males and females in the labor force for
2010 and 2011 have been used as an indicator in calculating the expected frequencies
for males and females in the restricted sample after deleting the outliers. The chi-
square test statistic is computed using the equation below
) ) eq. (4)
where is the observed frequency and is the expected frequency. The ratio of
women in the labor force was recorded as 42.57 percent in 2010 and 42.48 percent in
2011. Accordingly, the expected frequencies for 2010 are computed as 1,025.09 for
females and 1,329.91 for males which gave a chi-square value of 4.94 for 2010. On
the other hand, the expected frequencies have been calculated as 1022.92 for females
and 1385.08 for males yielding a chi- square test statistic of 5.34 for 2011. The
degrees of freedom is 1 for which the critical values at 0.02 level of significance are
5.412 and at 0.01 level of significance is 6.635. The test is conducted under the null
hypothesis that the population and sample proportions are same against the
alternative that they are not the same Therefore, for both years, the computed values
of the test statistics which are less than the critical values at both levels of
significance shows that one cannot reject the null hypothesis that the sample and
national proportions are same.
According to the test results, at the 0.01 level of significance, one cannot reject the
null hypothesis that the proportions are same. This would imply that the distribution
according to the observed values in the trimmed sample reflects the national level
proportions of males and females in the labor force. Our finding supports the
information provided on Figure 4 in Chapter 4 that outliers would distort the
31
estimations in the sense that the extreme values on the upper bound for males would
bias the estimates upward and those on the lower bound for females would bias the
estimates downward. This confirms that the trimmed data is appropriate to reflect the
actual proportions of males and females in the labor force.
4.2 Methodology
The econometric methodology applied in this research is cross-sectional Ordinary
Least Square (OLS) technique, based on Mincerian-type wage equation (Mincer,
1958 and 1974) which is logarithmic wages being a function of basically level of
education, experience and other variables represented by Z which may include
several variables such as gender, regional dummies, employment status categories
and geographical zones and various interaction terms as discussed in chapter two.
= ) eq. (5)
The regional dummy is identified as 1 for rural area and zero otherwise. The level of
employment categories are represented by binary dummy variables as self-employed
(ds), working for the government (dg) representing those employed in the public
sector and private sector (dp) to represent those employed in the private sector. The
geographical zone is made up of six geographical locations named as north west
(nw), north central (nc), north east (ne), south west (sw), south south (ss), and south
east (se).
32
Chapter 5
5 EMPIRICAL FINDINGS
This chapter will present empirical findings of Nigerian wage function based on
2010-2011 cross section data estimated by ordinary least squares (OLS) method.
Various specifications of the model are estimated starting with a simple model for
return to education and experience following Mincer (1974) but included gender
distinction. The model is then extended further by including regional, zonal and
employment status category variables in analyzing the determinants of wages in
Nigeria for the sample period.
5.1 The Basic Model: Return to Education and Experience
The estimated model is
= + + + + eq. (6)
where the dummy variable takes the value of 1 for males and 0 otherwise, education
variable (ed) is the number of years of schooling and X is the number of years of
experience as defined in chapter 4. The dependent variable is logarithmic value of
wages. The regression results are displayed in Table 7 below.
33
Table 7: Return to Education and Experience
Variable Coefficient Robust s.e t-stat. p-value 95% Conf. interval
constant
ed
X
Obs.
F(3,2403)
p-value
9.1347
0.0616
0.0034
-0.1595
2407
48.71
0.0000
0.0608
0.0886
0.0536
0.0021
0.0432
103.13
11.50
1.61
-3.70
0.000
0.000
0.107
0.000
8.9610 9.3084
0.0511 0.0721
-0.0007 0.0076
-0.2441 -0.0749
The annual return to education for females is 6.15 percent and is statistically highly
significant while return to experience is 0.35 percent and is marginally significant
only at 10 percent level of significance. On the other hand, the binary variable for
males is also highly significant and indicates that males earn 15.9 percent less than
females for the same level of education and experience. This finding is in line with
the descriptive statistics as shown by the box-plot in Figure 3 which shows females
earning more than males. Although the F statistics is highly significant, the
coefficient of determination is however 6.08 percent. From the above regression
results, education variable is found to play a major role in determining wages.
Moreover, the gender differential in wages is substantially large which requires
further investigation. In this respect, the model is extended to investigate whether
return to education is same for men and women given a constant wage differential as
evidenced in Table 7. Therefore, the male dummy variable is interacted with the
education variable. The estimation results are presented in Table 8.
34
Table 8: Return to Education by Gender
Variable Coefficient Robust
s.e
t-stat. p- value 95% Conf. interval
constant
ed
X
male
Obs.
F(4,2402)
p-value
9.0948
0.0658
0.0034
-0.0855
-0.0075
2407
36.63
0.0000
0.0611
0.1034
0.0078
0.0021
0.1070
0.0098
87.88
8.42
1.60
-0.80
-0.77
0.000
0.000
0.110
0.425
0.444
8.8919 9.2978
0.0505 0.0812
-0.0007 0.0076
-0.2955 0.1244
-0.0269 0.0118
The inclusion of the interaction term to the model resulted in insignificant coefficient
estimates for male dummy variable and the interaction term of male dummy with
education (maleed) which is due to the multicollinearity problem, as male and
maleed are highly correlated. Since the coefficient of male is the estimate of gender
differential when education is zero and since number of people with zero education is
only a few in the sample, it is more logical to estimate the gender differential at
average years of education. Therefore, the interaction term, instead, is calculated by
subtracting the average years of education which is about 10 years in the sample
Wooldridge, (2003). The estimated model is presented in Table 10 below where now
the coefficient for the male dummy variable has become highly significant indicating
that males earn 16.1 percent less than females. However, since the interaction term
(male with education) is not significant at the conventional levels, it means that as
education increases, wage differential between man and women does not change. In
this respect, the conclusion is that men earn less than women by 16.1 percent at all
levels of education which is the specification used in the extended models.
35
Table 9: Return to Education by Gender at Average Education Level
Variable Coefficient Robust s.e t-stat. p-value 95% Conf. interval
Constant
Ed
X
male
Obs.
F(4,2402)
p-value
9.0948
0.0658
0.0034
-0.1611
-0.0075
2407
36.63
0.0000
0.0611
0.1034
0.0078
0.0021
0.0431
0.0098
87.88
8.42
1.60
-3.74
-0.77
0.000
0.000
0.110
0.000
0.444
8.8919 9.2978
0.0505 0.0812
-0.0007 0.0076
-0.2457 -0.0766
-.0.0269 0.0118
One note is worth mentioning here that based on human capital theory diminishing
returns to experience and also to education are estimated by incorporating squared
values for experience and education respectively. However, the estimated
coefficients have been found to be negligibly small and statistically insignificant and
thus not reported.
5.2 Extended Model
After specifying the appropriate basic model, the variables regarding sectors as rural
versus urban, employment categories, as well as zones are added. The variable
named rural is a dummy variable given a value of 1 for those individuals living in the
rural area and 0 for individuals living in the urban area. The breakdown of the
employment can be summarized in six main groups as government (public sector),
self-employed, private sector including paid apprentice, international organization,
religious organization and others. As Table 6 in Chapter 4 indicates that most
individuals in the sample are self-employed, and the second highest share belongs to
the public sector and private sector comes as the third. Since remaining categories
are small the employment dummies are constructed for self-employed (ds)
36
government (dg) and private (dp). Zones are based on six geographical locations
named North West (ns), North Central (nc), North East (ne), South West (sw), South
South (ss), and South East (se). Moreover, one more variable is a dummy variable for
Christians (dc). Table 10 presents the results of the extended model.
37
Table 10: Extended Model compared with Basic Model
Independent variable Extended Model Basic Model
ed
0.0501
(0.0062)
[0.000]
0.0790
(0.0098)
[0.000]
X
0.0034
(0.0021)
[0.110]
0.0029
(0.0025)
[0.261]
dm
-0.1766
(0.0435)
[0.000]
-0.1858
(0.0522)
[0.000]
edmaleav
-0.0260
(0.0118)
[0.029]
dc
0.0901
(0.0678)
[0.184]
dg
0.0469
(0.0646)
[0.468]
ds
dp
-0.1558
(0.0510)
[0.002]
0.1044
(0.0741)
[0.159]
nc
-0.2043
(0.1093)
[0.062]
ne
-0.2167
(0.0661)
[0.001]
nw
-0.0754
(0.0852)
[0.376]
se
-0.0871
(0.0596)
[0.144]
ss
-0.2240
(0.0740)
[0.002]
rural
0.1016
(0.0545)
[0.062]
constant
9.2859
(0.1411)
[0.000]
9.0147
(0.1317)
[0.000]
Observations
R-squared
Ramsey
2403
0.0808
0.63[0.5980]
1937
0.0601
0.29[0.8316]
38
According to the estimation results, years of education, male dummy variable are
highly significant as before and the annual return to education is now 5 percent and
males earn 17.7 percent less than females. The dummy variable for self-employed is
also highly significant indicating that individuals who are self-employed most of
which are engaged in agriculture earn 15.6 percent less than others. The rural dummy
variable is significant at 6 percent level of significances and that those in the rural are
earn 10 percent higher than individuals in the urban area. North East and the South
South zone-dummy variables are highly significant with negative coefficient
estimates which show lowest wages earned in these zones, 21.6 percent and 22.4
percent less than the base category. The coefficient for experience variable is very
small which is 0.35 percent and is only significant at 11 percent. This finding means
that on-the-job training plays no significant role in wages determination which may
be due to the reason that most individuals are self-employed who work mostly in
rural areas. The p-value for the F statistics is zero which indicates that the variables
are jointly significant and the coefficient of determination increased to 8.1 percent as
compared to the basic model. The Ramsey test for omitted variables cannot reject the
null hypothesis of no omitted variables. The conclusion derived from the estimates
indicates the importance of education, gender and sector variables which leads to the
specification of a wage function for the rural sector. Table 11 presents the estimated
wage function model for the rural sector with and without the zone variables.
39
Table 11: Determinants of Wages for the Rural Sector
Independent
variable
Model 1 Model 2 Model 3
ed
0.0790
(0.0098)
[0.000]
0.0741
(0.0102)
[0.000]
0.0720
(0.0102)
[0.000]
X
0.0029
(0.0025)
[0.261]
0.0034
(0.0026)
[0.184]
0.0031
(0.0026)
[0.235]
male
-0.1858
(0.0522)
[0.000]
-0.2025
(0.0531)
[0.000]
-0.1903
(0.0525)
[0.000]
edmaleav
-0.0260
(0.0118)
[0.029]
-0.0228
(0.0117)
[0.052]
-0.0234
(0.0117)
[0.047]
dc
0.2291
(0.0636)
[0.000]
0.1516
(0.0811)
[0.062]
dg
-0.0532
(0.0731)
[0.466]
-0.0613
(0.0694)
[0.377]
ds
-0.1644
(0.0593)
[0.006]
-0.1846
(0.0585)
[0.002]
nc
-0.1891
(0.1346)
[0.160]
ne
-0.1669
(0.0790)
[0.035]
nw
-0.1225
(0.1075)
[0.255]
se
-0.1645
(0.0746)
[0.028]
ss
-0.1742
(0.0919)
[0.058]
dp
0.1013
(0.0829)
[0.222]
constant
9.0147
(0.1317)
[0.000]
8.9759
(0.1492)
[0.000]
9.1862
(0.1689)
[0.000]
Observations
R-squared
Ramsey
1937
0.0601
0.29[0.8316]
1934
0.0723
1934
0.0759
Basic model (Model 1), without zones (Model 2) and with zones (Model 3)
40
For the rural area first, the basic model is estimated where, however, also included
the interaction variable of male with education (edmaleav) which is highly
significant. All variables are significant except the experience variable. Regarding
the effect of education on log-wages is given by the first derivative of log-wages
with respect to the education variable as:
= 0.079 - 0.0260male
Accordingly, the marginal rate of return to education for males (when male = 1) is
5.3 percent while for females is 7.9 percent in the rural area that is larger than the
coefficient for the whole country as seen in Table 10. The coefficient of
determination for this regression function is 6.01 percent. The model is extended by
adding the employment categories where the employment in the government sector
and private sector are not significant, however, the binary variable for self-employed
is highly significant with a negative large coefficient indicating that the self-
employed earn 16.4 percent less than other employment categories in the rural area.
Another interesting point is that the dummy varibles for the religion has become
highly significant with a large positive coefficient meaning that in the rural area,
christians earn 23 percent higher than the base category. The coefficient of
determinion has increased from 6.01 percent to 7.23 percent by inclusion of the
employment categories and relion variable. When the zone variables are added, the
regression coefficients and their statistical significance do not change much while
only two of the zone binary variables, North Central (nc) and North West (nw) are
not statistically significant. However, the coefficient of determination only slightly
increased to 7.6 percent from 7.2 percent by the addition of the zone variables.
41
The estimation results indicate that years of scooling representing various levels of
education has a substantial role in wage determination in comparison to employment
category, zones and rural versus urban sectors. Gender gap is another important
factor for the Nigerian wage function where males earn significantly lower than
females. However, the wage differential betwen men and women is constant which
means that men earn less than females at all levels of education for the wage function
estimated countrywise whereas, the return to education is not same for males and
females at all levels of education for the rural area. In Table 11, the interaction term,
male.ed is highly significant and negative in all the models (basic model and the
extended models) indicating that the gap increases as the education gets larger in the
rural area. This may indicate negative return to education in the rural area. However,
in all the regressions, the coefficient of determination is between about 6 percent and
8 percent which is rather low. This indicates that in the Nigerian labor market there
are other factors influencing wages rather than personal characteristics and
employment and geographical factors. This finding supports the report of Aminu
(2011) that Nigerian government wage review policies, ad hoc wage commissions
play an important role in wage determination rather than demand and supply forces
in the labor market.
42
Chapter 6
6 SUMMARY AND CONCLUSION
This chapter shall present the general overview of this research work, conclusions
and remarks. Further we will itemize some policy suggestions including suggestions
for further research in this area.
6.1 Conclusions
So far, the main aim of this study is to investigate and empirically examine the
various determinants of wages using a cross-sectional data for Nigeria. In particular,
this study was targeted towards individual households across the regions of Nigeria
and within the agricultural sector. Through this research, we were able to contribute
to the literature on labour market analysis for a typical developing country such as
Nigeria. Hence our lick-off point in the empirical analysis is the introduction of the
Mincerian-type wage equation of Mincer (1974) incorporating individual attributes,
region, gender and other employment categories in the analysis.
Considering the return to education and experience, our findings have shown that
females earn quite higher than the males even across the same level of education and
experience. Even when the model was further extended so as to investigate if
earnings across gender would change at constant wages, our observation also
presents similar results. Also while including both regional and employment
characteristics of individuals, males were observed to earn about 17percent lower
43
than female folks. In the regional basis, both North East and South-South zones
showed less in earnings compared to the base category and on-the-job training plays
little or no significance on wages. Similarly by including sectorial employment
status, we have been able to show that self-employed individuals earn lower than
other employment status in the rural areas.
Overall, education which is one of the basic components of human capital theory has
a positive effect in wage determination in rural Nigeria. These findings correlate well
with other similar findings for Nigeria such as Aminu (2010 and 2011), and
Ogwumike (2006).
It is our view that the Nigerian labour market is heterogeneous in nature due to its
large population including multi-ethnic and religious orientations. However,
education which has been made free especially at the primary and secondary levels is
indeed one of the crucial elements of wage determination across the regions of the
country. Since education is productivity augmenting, major policy frameworks
should actually be directed towards implementing various educational and job-
training programs to encourage productivity especially within the informal
agricultural sectors of Nigeria. It is also surprising that the South-South part of
Nigeria which is the beehive of the Nigeria’s rich oil resources has shown lower
earnings than other categories, in the region.
6.2 Recommendations for Further Studies
It has become obvious that the gender differential in wages appears to be quite large,
further studies should possibly be directed towards determining the actual cause of
such differences in wages and comparing such findings with the next wave of the
44
data. Similarly, regional wage differences should studied especially within the South-
South regions of Nigeria. It is expected that this region should show a considerable
high wage structure given their contributions to the Nations’ economy. Finally, it
will usher in a new and a strong comparative stance of the labour market of Nigeria,
if a panel study is carried out using the most recent wave of the GHS data.
45
REFERENCES
Aluede R.O.A. (2006), 'Universal Basic Education in Nigeria: Matters Arising',
Journal of Human Ecology, 20(2): 97-101 (2006)
Amaghionyeodiwe, L.A. and Osinubi, T.S. (2006), 'The Nigerian Educational
System and Return to Education', International Journal of Applied Econometrics and
Quantitative Studies Vol.3-1
Aminu, A. (2007a), ‘Were Public Sector Workers Still Underpaid in Nigeria after the
Wage Increase in May 2000?’, Ibadan Journal of Social Science, Volume 5, Number
1:55-72.
Aminu, A. (2007b), ‘Government Wage Review Policy and Public-Private Sector
Wage Differential in Nigeria’, A Revised Final Report Submitted to the African
Economic Research Consortium (AERC), Nairobi, Kenya.
Aminu, A. (2011), ‘Determinants of Participation and Earnings in Wage
Employment in Nigeria’, 5th IZA/World Bank Conference: Employment and
Development
Analogbei, F.C.O. (2000).’Trade Reforms and Productivity in Nigeria’, Central
Bank, www.cenbank.org/out/publications/ occasionalpapers/rd/2000/abe-00-9.pdf
46
Aromolaran, A.B. (2006), ‘Estimates of Mincerian Returns to Schooling in Nigeria’,
Oxford Development Studies, 34:2, 265-292
Arrow, K. J. (1973). ‘Higher education as a filter’, Journal of Public Economics,
Elsevier, vol. 2(3), pages 193-216.
Becker, G. (1962). "Investment in Human Capital: A Theoretical Analysis," Journal
of Political Economy, University of Chicago Press, vol. 70, pp 9
Becker, G.S. (1964, 1993, 3rd ed.). 'Human Capital: A Theoretical and Empirical
Analysis, with Special Reference to Education'. Chicago, University of Chicago
Press. 978-0-226-04120-9
Bedi, A.S. (1998). ‘Sector choice, multiple job holding and wage differentials:
Evidence from Poland’. The Journal of Development Studies, 35(1): 162–79.
Bennel, P.S. (1981), ‘Earnings Differentials between Public and Private Sectors in
Africa: The Cases of Ghana, Kenya, and Nigeria’. Labour and Society, 6(3): 223-41.
Berndt, E.R. (1991). ‘The Practice of Econometrics’, Addison Wesley, 1991 Ch 5 pp
154
Boudarbat, B. (2004), ' Employment Sector Choice in a Developing Labor Marrket',
Department of Economics, University of British Columbia
47
Card, D. (1993), ‘Using geographic variation in college proximity to estimate the
return to Schooling’, National Bureau of Economic Research Working Paper No.
4483
Central Bank of Nigeria (2010) Statistical Bulletin Online Data base,
www.cenbank.org/documents/data.asp
Corbo, V. and Stelcner, M. (1983). ‘Earnings Determination and Labour Markets:
Gran Santiago’, Chile - 1978. J. Dev. Econ 12:251-266
Edho, O.G. (2009). 'The Challenge Affecting the Implementation of the Universal
Basic Education (UBE) in Delta State'.Journal of social sciences Vol 20, No. 3, pp 1
Fabusoro, E., Omotayo, A., Apantaku, S. O. & Okuneye, P. A. (2010), ‘Forms and
Determinants of Rural Livelihoods Diversification in Ogun State, Nigeria’, Journal
of Sustainable Agriculture, 34:4, 417-438.
Federal Office of Statistics (2001), ‘General household survey’, 1999, FOS, Lagos
Federal Republic of Nigeria (2000), ‘Obansanjo’s Economic Direction 1999-2003’,
Abuja: Federal Government Press
Filmer, D. and Lindauer, D.L (2002), 'Does Indonesia Have a "Low Pay" Civil
Service?', Policy Reasearch Working Paper, Development Research Group, World
Bank Vol. 2621
48
Griliches and Mason (1972), ‘Education, Income and Ability’, Journal of Political
Economy 80 (3): P 74-103
Hanoch, G. (1967), ‘An Economic Analysis of Earning and Schooling’, Journal of
Human Resources, vol 2, no 3, p 310-29
Hyder, A. and Reilly, B. (2005), ‘The public sector pay gap in Pakistan: A Quantile
Regression Analysis’, PRUS Working Paper No. 33. Poverty Research Unit at
Sussex, University of Sussex, Brighton
House R.J (1986), 'Organizational Legitimacy and the Liability of Newness',
Administrative Science Quarterly, Vol. 31, No. 2, pp. 171-193.
Ikhizama B. O. and Lawal B. O. (2006), ‘Women Scientists' Involvement in
Agricultural Development in Southwest Nigeria’, Journal of Agricultural & Food
Information, 7:4, 11-22.
Jonah A. and Yousuo (2013), ‘The Impact of Wage Differentials on Labour Turn
Over in Nigeria’, International Journal of Academic Research in Business and Social
Sciences, Vol. 3, No. 9
Kerr, C., Dunlop, J.T., Harbison, F.H and Meyers, C.A. (1962).’ Industrialism and
Industrial Man’, London: Heinemann.
Lam. D and Schoeni R.F. (1993), ‘Effect of family background on Earning and
Return to Schooling in Brazil’. Journal of Political Economy 101(4): 710-40
49
Leontarridi M.R. (1998). ‘Segmented labour marketers theory and evidence’. Journal
of economic surveys 12(1) p 63-101
Lindauer, D.L. and Sabot R. (1983), 'The Public/Private Wage Differential in a Poor
Urban Economy', Journal of Development Economics, Vol. 12.
Lindauer, D.L, Meesook O.A, and Suebsaeng P. (1988), ‘Government Wage Policy
in Africa: Some Findings and Policy Issues’, World Bank Research Observer, Vol.3,
No.1: 1-23.
Matthew A. O. and Azuh D. (2013), ‘Trade Liberalization and Formal-informal
Sector Wage Differential in Nigeria’, Asian Economic and Financial Review, 2013,
3(12):1651-1669.
Mincer, J. (1958), ‘Investment in Human Capital and Personal Income Distribution’,
Journal of Political Economy, 56(4): 281–302.
Mincer. J. (1962). ‘Labour Force Participation of Married Women in Aspect of
Labour Economics’, Princeton: Princeton University Press, for NBER ch 1, Vol 2
Mincer, J. (1974), ‘Schooling, Experience and Earnings’, National Bureau of
Economic Research, New York: Columbia University Press.
Mincer, J. and Polachek S. (1974). ‘Family investment in human capital: Earning of
women’. Journal of Political Economy, 82(2)
50
Mohan, R. (1986), ‘Work, Wages, and Welfare in a Developing Metropolis:
Consequences of Growth in Bogota, Columbia’, New York: Oxford University Press.
Nigerian Bureau of Statistics (2010), General household survey, 2010, FOS, Lagos,
http://www.nigerianstat.gov.ng/
Nwabugo, N.E. (2011), ‘The Story of Structural Adjustment Programme in Nigeria
from the Perspective of the Organized Labour’, Australian Journal of Business and
Management Research Vol.1 No.7 pp 30-41
Ogundari, K. and Aromolaran, A.B. (2014), ‘Impact of Education on Household
Welfare in Nigeria’, International Economic Journal, 28:2, 345-364
Ogwumike, F. O., Adubi, A. A. and Agba, V. A. (2002), ‘Major sources and levels
of risk in Nigeria, Report of a pilot study on social protection strategy for Nigeria’,
Nigeria: Department of Economics, University of Ibadan. Nigeria
Ogwumike, F.O., Alaba O.A., Alaba O.B., Alayande B.A., Okojie C.A.A. (2006),
‘Labour Force Participation, Earnings and Inequality in Nigeria’, African Economic
Research Consortium (AERC), vol. 3, pages 29-33.
Okigbo, B.N. (1991), ‘Development of sustainable agricultural system in Africa’,
Distinguished African Scientist Lecture Ser. 1. IITA, Ibadan, Nigeria
Psacharopoulos, G. (1985). ‘Returns to Education: A Further International Update
and Implications’. The Journal of Human Resources, vol. XX, no. 4,pp. 584-604.
51
Psacharopolous, G. and Patrinos, H.A. (2002), 'Returns to Investment in Education:
A Further Update', World Bank Policy Research Working Paper 2881
Samarrai, S. and Reilly, B. (2005), 'Education, Employment and Earnings of
Secondary School-Leavers in Tanzania: Evidence from a Tracer Study'. PRUS
Working Papers 31, Poverty Research Unit at Sussex, University of Sussex.
Schultz, P.T. (2001), 'Women's roles in the agricultural household: Bargaining and
human capital investments', Handbook of Agricultural Economics, in: B. L. Gardner
& G. C. Rausser (ed.), Handbook of Agricultural Economics, edition 1, volume 1,
chapter 8, pages 383-456 Elsevier.
Spence, A. M., (1973), ‘Time and Communication in Economic and Social
Interaction’, The Quarterly Journal of Economics, MIT Press, vol. 87(4), pp 651-60
Spence, A.M, (1973), ‘Job Market Signaling’, The Quarterly Journal of Economics,
MIT Press, vol. 87(3), pages 355-740
Spence, A.M, (1974). ‘Competitive and optimal responses to signals: An analysis of
efficiency and distribution’, Journal of Economic Theory, Elsevier, vol. 7(3), pages
296-332
Steier, F. (1987), ‘Schooling, Experience, and Earnings: Issues in Venezuelan
Development’, 1975–84”. PhD. dissertation, Columbia University, New York
52
Taubman, P. (1976a), ‘Earning, Education, Genetics, and Environment’, Journal of
Human Resources, 11, 447-61
Taubman, P. (1976b), ‘The Determinants of Earning: Genetics, Family, and other
Environments; a study of male white twins’, American Economic Review, 66, 5, Dec
1976, 858-70.
Temesgen, T. (2008), ‘Effects of Labor Market Institutions and Establishment
Characteristics on Gender Wage Inequality in Africa’, International Journal of
Sociology and Social Policy, Vol. 28 Iss 9/10 pp. 309 - 325.
Terrell, K. (1993), 'Public-Private Wage Differentials in Haiti: Do Public Servants
Earn a Rent?', Journal of Development Economics, 42(2): 293-314.
Thurow, L.C. (1975), 'Generating inequality: mechanisms of distribution in the U.S.
Economy', New York: Basic Books.
Tokarick, S. (2008), ‘Dispelling some misconception about Agricultural Trade
Liberalization’. Journal of Economic Perspectives, 22, No. 1, pp 199-216
Ugochukwu A. and Chijioke J. E. (2011), ‘Macroeconomic Policy for full and
Productive and Decent Employment for all: the case of Nigeria’, International
Labour Office, Employment Policy Department. - Geneva: ILO, 2011 p 65, No. 107
UNDP (2009), ‘United Nations Development Program Human Development
Reports’, http://hdr.undp.org/en/content/human-development-report-2009
53
US Embassy (2012) Retrieved 16/03/2014. Nigerian Education fact sheet, www.
http://photos.state.gov
Van der Gaag, J., M. Stelcner and W. Vijverberg. (1989), ‘Wage Differentials and
Moonlighting by Civil Servants: Evidence from Cote d’Ivoire and Peru’, The World
Bank Economic Review,3:(1): 67–95.
Wooldridge, J.M. (2003). ‘Introductory Econometrics: A Modern Approach’, 5th
Edition, P/233
World Bank (2010), ‘Globalisation, Growth and Poverty’, World Bank
World Bank Development Indicators, (2011), Online Data base.